Transformer iron core vibration monitoring and fault diagnosis method based on voiceprint features

Abstract

The invention discloses a transformer iron core vibration monitoring and fault diagnosis method based on voiceprint characteristics, and belongs to the technical field of power equipment monitoring and fault diagnosis, and the method comprises the steps: obtaining voiceprint signals of each monitoring point about the vibration of an iron core in a preset transformer, and collecting the voiceprint characteristics according to the voiceprint signals of different monitoring points, analyzing the change rule of each voiceprint signal; according to the change rule, distinguishing the change of the voiceprint signal according to light change and moderate change, and obtaining the operation state of the preset transformer under the condition of light change; the operation state comprises current parameters of a preset transformer; and giving a monitoring period according to the light change, and presetting a critical threshold value related to the vibration frequency change of the voiceprint signal in different monitoring periods. According to the invention, by monitoring and analyzing the voiceprint signal and combining the technical means of multiple monitoring points, dynamic monitoring and enhanced learning, early fault early warning and accurate diagnosis of the operation state of the transformer are realized, and the method has relatively high practicability and innovativeness.

Description

Technical Field

[0001] The present invention relates to the technical field of power equipment monitoring and fault diagnosis, and in particular to a transformer core vibration monitoring and fault diagnosis method based on voiceprint features. Background Art

[0002] The transformer is the core equipment in the power system, mainly composed of iron core, winding, structural fasteners and insulating components. The operating status of the transformer is directly related to the safety and stability of the power grid. The iron core is the key component of the transformer, and its vibration characteristics are closely related to the operating status.

[0003] The main reason for the vibration of the transformer core is that the silicon steel sheet undergoes periodic expansion and contraction vibration due to its magnetostrictive characteristics under the influence of the alternating magnetic field. The transformer fault state caused in this case is usually accompanied by changes in the electromagnetic field, which in turn affects the changes in the vibration and sound of each device component of the transformer.

[0004] Traditional transformer fault diagnosis methods mainly rely on electrical parameter monitoring or oil chromatography analysis, but these methods have problems such as poor real-time performance and inability to provide early warning. In recent years, voiceprint recognition technology has gradually been applied in the industrial field. By analyzing the sound signals generated when the equipment is running, non-contact monitoring and fault diagnosis of the equipment status can be achieved. Among them, the application document with the technical application number CN202310068167.0 provides a transformer fault detection method, system and medium based on voiceprint features. The technical solution includes collecting voiceprint images of transformers; extracting voiceprint contours and transformer contours from the voiceprint images using a trained mask convolutional neural network model; calculating the relative intersection area I between the voiceprint contour and the transformer contour. If the relative intersection area I is greater than the set value Ith, it is determined that the transformer is faulty, and the transformer image is cut out from the voiceprint image according to the transformer contour. The technical solution can realize non-intrusive live detection of the transformer operating status and reduce the impact of the complex substation environment.

[0005] Another application document with the technical application number CN202410952483.9 provides a transformer status diagnosis method, equipment, medium and product based on voiceprint, which includes: selecting key points based on vibration sensitive points and vulnerable points to obtain monitoring key point information, and the monitoring key point information provides an installation location for the installation of the voiceprint sensor; then, obtaining the voiceprint signal collected by the voiceprint sensor installed at the key point, filtering the voiceprint signal, and performing filtering to remove environmental noise in the voiceprint signal to improve the signal-to-noise ratio of the voiceprint signal. This technical solution sets the voiceprint sensor at these key points, which can capture changes in the operating status of the transformer and improve the effectiveness of voiceprint monitoring.

[0006] However, in the prior art, in the case of minor transformer faults (such as slight looseness of the core) or early fault stages, the characteristics of the soundprint signal are not obvious enough and difficult to distinguish from the normal operating state, resulting in the problem of low accuracy in soundprint monitoring and fault diagnosis of transformer core vibration, which limits its widespread application in practice. Summary of the invention

[0007] In view of the above-mentioned problems existing in the technical field of existing power equipment monitoring and fault diagnosis, the present invention is proposed.

[0008] Therefore, one of the objects of the present invention is to provide a transformer core vibration monitoring and fault diagnosis method based on voiceprint features, which utilizes the monitoring and analysis of voiceprint signals, combined with multiple monitoring points, dynamic monitoring, and enhanced learning techniques, to achieve early fault warning and accurate diagnosis of the transformer operating status, and has high practicality and innovation.

[0009] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0010] The present invention provides a transformer core vibration monitoring and fault diagnosis method based on voiceprint features, comprising the following steps:

[0011] Step S10: obtaining the voiceprint signal of the core vibration at each monitoring point in the preset transformer, collecting voiceprint features according to the voiceprint signals at different monitoring points, and analyzing the change rules of each voiceprint signal;

[0012] Step S20: according to the change rule, the voiceprint signal change is distinguished into a slight change and a moderate change, and the operation state of the preset transformer in the case of the slight change is obtained; the operation state includes collecting the current parameters of the preset transformer;

[0013] Step S30: According to the slight change given monitoring period, a critical threshold value of the change of the vibration frequency of the voiceprint signal is preset in different monitoring periods, and before the critical threshold value, the collected change of the vibration frequency of the voiceprint signal is divided into in, represents the vibration frequency of the nth voiceprint signal. In a certain monitoring period, if the vibration frequency of the voiceprint signal changes toward the critical threshold, it is determined that the change of the voiceprint signal evolves from a slight change to a moderate change, and a fault warning is issued for the preset transformer;

[0014] Step S40: Based on the evolution law, analyzing the monitoring period in which the vibration frequency of the voiceprint signal with the most occurrences changes toward the critical threshold, and analyzing the change trend of the current parameter in the monitoring period;

[0015] Step S50: presetting a safety threshold according to the change trend of the current parameter, and during the monitoring period, if the change of the current parameter does not exceed the safety threshold, then canceling the fault warning; otherwise, then not canceling it.

[0016] As a preferred solution of the present invention, in step S10, the monitoring points in the preset transformer include the surface of the core column, the inner wall of the transformer oil tank, the winding insulation pressure plate, the surface of the transformer oil tank, the insulation layer between the core and the winding, the bottom or top of the transformer oil tank and the core grounding lead.

[0017] As a preferred solution of the present invention, in step S30, according to the monitoring cycle given by the slight change, including monitoring with each natural day as a monitoring cycle, a set of initial vibration frequencies of the voiceprint signal is obtained in different monitoring cycles, and the number of the initial vibration frequencies is at least 3 to 5; the current parameters corresponding to each of the initial vibration frequencies are collected, and the current parameters are analyzed and compared according to the initial vibration frequencies obtained in different monitoring cycles, and the difference effect of different initial vibration frequencies on the current parameters is calculated, and the calculation is obtained according to the following formula:

[0018] Among them, p j represents the pth initial vibration frequency obtained in the jth monitoring period;

[0019] In the formula, k j represents the kth maximum current parameter collected in the jth monitoring cycle, x represents the average vibration frequency obtained from the initial vibration frequencies of the same group of voiceprint signals, h represents the difference in the initial vibration frequencies of the same group of voiceprint signals, and t represents the time required for adjacent differences to change;

[0020] Within the time period, when the change of the current parameter is faster than the change of the adjacent difference, it is determined that the preset transformer is in an abnormal operating state and a fault warning is issued; otherwise, no determination is made.

[0021] As a preferred solution of the present invention, when it is determined that the preset transformer is in an abnormal operating state, a monitoring duration is given in the corresponding monitoring cycle, and the monitoring duration includes continuous monitoring for 10 to 15 minutes; 3 to 6 frequency values ​​with the highest vibration frequency are collected during the monitoring duration, and the frequency values ​​are marked as reference values; and the current parameter change is obtained based on the smallest frequency value, and when the current parameter does not increase with the increase of the frequency value, the fault warning is lifted; otherwise, the fault warning is maintained.

[0022] As a preferred solution of the present invention, wherein: the changing trend of the average vibration frequency is collected in different monitoring cycles of each of the monitoring points, and a data set is generated; at least 3 monitoring points that change towards an increasing trend are obtained in the data set, and the maximum current parameters of the corresponding monitoring points when the increasing trend changes are analyzed, and at the same time, a risk threshold is preset according to the average vibration frequency corresponding to the maximum current parameter; when the changing trend of the average vibration frequency of other monitoring points of the preset transformer in the future time period is greater than that of the 3 monitoring points, it is determined that the preset transformer is in an abnormal operating state, and a fault warning is issued; otherwise, no determination is made.

[0023] As a preferred solution of the present invention, when it is not determined that the preset transformer is in an abnormal operating state, an acquisition time is preset at the corresponding monitoring point, and the acquisition time includes one acquisition time per minute; within the acquisition time, the time required for each increase of 1 Hz frequency is obtained based on the average vibration frequency corresponding to the maximum current parameter. When in any of the monitoring cycles, if the time shows a shortening trend, it is determined that the operating state of the preset transformer is in an abnormal change state, and a fault warning is issued; otherwise, no determination is made.

[0024] As a preferred solution of the present invention, the Q-learning algorithm is used in the data set for reinforcement learning, wherein the reinforcement learning includes giving at least three target monitoring nodes in any monitoring cycle, the three target monitoring nodes include a morning node, a noon node and an evening node, and calculating the proportion of the vibration frequency of each target monitoring node to the vibration frequency of the corresponding monitoring cycle, which is calculated according to the following formula:

[0025] Among them, w i represents the wth maximum vibration frequency obtained for the i-th time in the monitoring period;

[0026] Where y represents the maximum vibration frequency obtained at each target monitoring node, m represents the average vibration frequency obtained at each target monitoring node, (l i ) represents the lth vibration frequency difference collected at each target monitoring node for the i-th time;

[0027] According to the preset proportion value of the vibration frequency of each target monitoring node in the corresponding monitoring period, when the vibration frequency of each target monitoring node is lower than the proportion value, it is determined that the operating state of the preset transformer is in an abnormal change state and a fault warning is issued; otherwise, no determination is made.

[0028] As a preferred solution of the present invention, the changing pattern of the vibration frequency is analyzed at each target monitoring node, each vibration frequency is divided into a number of evaluation indicators, and the median vibration frequency of each target monitoring node is statistically calculated according to the evaluation indicators. When the median vibration frequency shows a decreasing trend, it is determined that the operating state of the preset transformer is in an abnormal change state, and a fault warning is issued; otherwise, no determination is made.

[0029] A terminal, characterized in that it includes a processor, an input interface, an output interface and a memory, wherein the processor, input interface, output interface and memory are interconnected, wherein the memory is used to store a computer program, the computer program includes program instructions, and the processor is configured to call the program instructions to execute the method as described above.

[0030] A computer-readable storage medium is characterized in that the computer-readable storage medium stores a computer program, wherein the computer program includes program instructions, and when the program instructions are executed by a processor, the processor executes the method as described above.

[0031] 1. The present invention monitors the change law of the soundprint signal of the core vibration, divides the change of the soundprint signal into slight change and moderate change, sets the critical threshold, identifies the change trend of the fault in advance, and realizes the early detection and diagnosis of the preset transformer fault;

[0032] 2. By collecting voiceprint signals at multiple key locations of the preset transformer (such as the surface of the core column, the inner wall of the oil tank, the winding insulation plate, etc.), and combining them with current parameters for comprehensive analysis, this multi-monitoring point and multi-dimensional analysis method can more comprehensively reflect the operating status of the transformer and improve the accuracy of fault diagnosis;

[0033] 3. By taking natural days as the monitoring period and setting the critical thresholds of the initial vibration frequency and current parameters within the monitoring period, through continuous monitoring (such as 10 to 15 minutes) and real-time data analysis, the changes in the voiceprint signal can be dynamically monitored, and the fault warning status can be adjusted in real time according to the changing trend of the current parameters. This dynamic monitoring mechanism can timely capture the changes in the preset transformer operating status and reduce misjudgment. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor. Among them:

[0035] Figure 1A schematic diagram of a method flow chart of an embodiment of the present invention;

[0036] Figure 2 It is a schematic diagram of the process structure of an embodiment of the present invention. DETAILED DESCRIPTION

[0037] In order to make the purpose, technical solution and advantages of the embodiment of the present invention clearer, the technical solution of the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings of the embodiment of the present invention. Obviously, the described embodiment is a part of the embodiment of the present invention, not all of the embodiments. Based on the described embodiment of the present invention, all other embodiments obtained by ordinary technicians in this field belong to the scope of protection of the present invention.

[0038] In the prior art, during a minor transformer fault (such as a slightly loose core) or early fault stage, the characteristics of the soundprint signal are not obvious enough to distinguish it from the normal operating state. As a result, the soundprint monitoring and fault diagnosis of transformer core vibration still have the problem of low accuracy, which limits its widespread application in practice.

[0039] Based on this, the present invention proposes a transformer core vibration monitoring and fault diagnosis method based on voiceprint features, which utilizes the monitoring and analysis of voiceprint signals, combined with multiple monitoring points, dynamic monitoring, and enhanced learning techniques, to achieve early fault warning and accurate diagnosis of the transformer operating status, and is highly practical and innovative.

[0040] The present invention is further described in detail below through embodiments and in conjunction with the accompanying drawings.

[0041] Reference Figure 1 to Figure 2 , is an embodiment of the present invention, which provides a transformer core vibration monitoring and fault diagnosis method based on voiceprint features, comprising the following steps:

[0042] Step S10: obtaining the voiceprint signal of the core vibration at each monitoring point in the preset transformer, collecting voiceprint features according to the voiceprint signals at different monitoring points, and analyzing the change rules of each voiceprint signal;

[0043] It should be noted that in this embodiment, the monitoring points in the transformer are preset, including the surface of the core column, the inner wall of the transformer oil tank, the winding insulation pressure plate, the surface of the transformer oil tank, the insulation layer between the core and the winding, the bottom or top of the transformer oil tank and the core grounding lead;

[0044] The core vibration mainly comes from the magnetostrictive effect. Its vibration signal is most direct and obvious on the surface of the core column. Installing the sensor on the surface of the core column can effectively capture the normal, longitudinal and lateral vibration signals of the core.

[0045] The inner wall of the transformer oil tank, the core vibration signal is transmitted to the inner wall of the oil tank through the transformer oil. The vibration signal of the inner wall of the oil tank can reflect the vibration characteristics of the core. In addition, the signal of the inner wall of the oil tank is relatively stable and is less affected by external interference;

[0046] Winding insulation plate, the core vibration signal will be transmitted to the insulation plate through the winding. The signal here can reflect the vibration coupling between the core and the winding. At the same time, the vibration signal on the winding insulation plate can be compared and analyzed with the core vibration signal to separate the vibration of the winding and the core.

[0047] The surface of the transformer oil tank. Although the signal on the surface of the oil tank is a mixed signal of the core and winding vibration, the vibration characteristics of the core can be extracted through appropriate signal separation technology. In addition, the signal acquisition on the surface of the oil tank is relatively simple and suitable for field application.

[0048] The insulation layer between the core and the winding. Before the core vibration signal is transmitted to the winding, it will produce an obvious vibration response in the insulation layer. The signal here can directly reflect the vibration state of the core while avoiding the interference of the winding vibration.

[0049] At the bottom or top of the transformer tank, the core vibration signal is transmitted to the bottom or top of the tank through the transformer oil. The signal here can reflect the overall vibration state of the core. At the same time, the signal at the bottom or top of the tank is relatively stable and less affected by external interference.

[0050] Near the core grounding lead, weak vibration signals will be generated near the core grounding lead. These signals are closely related to the vibration state of the core. By monitoring the signals near the grounding lead, the vibration state of the core can be indirectly reflected;

[0051] Step S20: according to the change rule, the voiceprint signal change is distinguished into slight change and moderate change, and the operation state of the preset transformer in the case of slight change is obtained; the operation state includes collecting the current parameters of the preset transformer;

[0052] Step S30: According to the given monitoring period of slight changes, a critical threshold value of the change of the vibration frequency of the voiceprint signal is preset in different monitoring periods, and before the critical threshold value, the collected vibration frequency change of the voiceprint signal is divided into in, Indicates the vibration frequency of the nth voiceprint signal. In a certain monitoring period, if the vibration frequency of the voiceprint signal changes toward the critical threshold, it is determined that the change of the voiceprint signal evolves from a slight change to a moderate change, and a fault warning is issued for the preset transformer;

[0053] It should be noted that in this embodiment, a monitoring cycle given according to slight changes includes monitoring with each natural day as a monitoring cycle, obtaining a set of initial vibration frequencies of the voiceprint signal in different monitoring cycles, and the number of initial vibration frequencies is at least 3 to 5; collecting current parameters corresponding to each initial vibration frequency, and analyzing and comparing the current parameters according to the initial vibration frequencies obtained in different monitoring cycles, calculating the difference effect of different initial vibration frequencies on the current parameters, and calculating according to the following formula:

[0054] Among them, p j represents the pth initial vibration frequency obtained in the jth monitoring period;

[0055] In the formula, k j represents the kth maximum current parameter collected in the jth monitoring cycle, x represents the average vibration frequency obtained from the initial vibration frequencies of the same group of voiceprint signals, h represents the difference in the initial vibration frequencies of the same group of voiceprint signals, and t represents the time required for adjacent differences to change;

[0056] Within the time length, when the change of the current parameter is faster than the change of the adjacent difference, it is determined that the preset transformer is in an abnormal operating state and a fault warning is issued; otherwise, no determination is made;

[0057] It should be noted in this embodiment that when the transformer is in stable operation, its current is usually relatively stable, but it changes regularly according to the change of load, rather than changing rapidly or drastically. Therefore, it is of practical significance to calculate the difference effect of different initial vibration frequencies on the current parameters.

[0058] Step S40: Based on the evolution law, analyzing the monitoring period in which the vibration frequency of the voiceprint signal with the most occurrences changes toward the critical threshold, and analyzing the change trend of the current parameter in the monitoring period;

[0059] Step S50: Preset a safety threshold according to the change trend of the current parameter. During the monitoring period, if the change of the current parameter does not exceed the safety threshold, the fault warning is released; otherwise, it is not released;

[0060] Further, in this embodiment, when it is determined that the preset transformer is in an abnormal operating state, a monitoring duration is given in the corresponding monitoring cycle, and the monitoring duration includes continuous monitoring for 10 to 15 minutes; 3 to 6 frequency values ​​with the highest vibration frequency are collected during the monitoring duration, and the frequency values ​​are marked as reference values; and the current parameter change is obtained based on the smallest frequency value, and when the current parameter does not increase with the increase of the frequency value, the fault warning is released; otherwise, the fault warning is maintained;

[0061] On the basis of the above, the present embodiment collects the changing trend of the average vibration frequency in different monitoring cycles of each monitoring point and generates a data set; obtains at least 3 monitoring points that change towards an increasing trend in the data set, and analyzes the maximum current parameters of the corresponding monitoring points when the increasing trend changes, and presets the risk threshold according to the average vibration frequency corresponding to the maximum current parameter; when the changing trend of the average vibration frequency of other monitoring points of the preset transformer in the future time period is greater than that of the 3 monitoring points; it is determined that the preset transformer is in an abnormal operating state and a fault warning is issued; otherwise, no determination is made.

[0062] Furthermore, in this embodiment, when it is not determined that the preset transformer is in an abnormal operating state, an acquisition time is preset at the corresponding monitoring point, and the acquisition time includes one acquisition time per minute; within the acquisition time, the time required for each increase of 1 Hz frequency is obtained based on the average vibration frequency corresponding to the maximum current parameter. When in any monitoring cycle, if the time shows a shortening trend, it is determined that the operating state of the preset transformer is in an abnormal change state, and a fault warning is issued; otherwise, no determination is made;

[0063] On the basis of the above, it is important to emphasize in this embodiment that the Q-learning algorithm is used in the data set for reinforcement learning, wherein the reinforcement learning includes giving at least three target monitoring nodes in any monitoring cycle, the three target monitoring nodes including a morning node, a noon node and an evening node, and calculating the proportion of the vibration frequency of each target monitoring node to the vibration frequency of the corresponding monitoring cycle, which is calculated according to the following formula:

[0064] Among them, w i represents the wth maximum vibration frequency obtained for the i-th time in the monitoring period;

[0065] Where y represents the maximum vibration frequency obtained at each target monitoring node, m represents the average vibration frequency obtained at each target monitoring node, (l i ) represents the lth vibration frequency difference collected at each target monitoring node for the i-th time;

[0066] According to the preset weight value of each target monitoring node in the corresponding monitoring period vibration frequency, when the vibration frequency of each target monitoring node is lower than the weight value, it is determined that the operation state of the preset transformer is in an abnormal change state, and a fault warning is issued; otherwise, no judgment is made;

[0067] It should be noted that in this embodiment, noon is usually the peak period of electricity consumption, which is mainly affected by industrial, commercial and residential electricity loads. The load current of the transformer increases significantly. In this case, the current changes frequently, which will increase the vibration frequency.

[0068] Evening is also the peak period of electricity consumption. However, as time goes by, some industrial loads gradually decrease, and residential electricity loads may increase. The change of current is more complicated, which may include the start and stop of lighting, air conditioning and other equipment. In the early evening (such as 18:00-22:00), the current is large and the vibration frequency of the iron core is high.

[0069] Therefore, this embodiment determines that the operation state of the preset transformer is in an abnormal change state based on the vibration frequency of each target monitoring node being lower than the specific gravity value, which has practical significance;

[0070] At the same time, this embodiment analyzes the change law of the vibration frequency at each target monitoring node, divides each vibration frequency into several evaluation indicators, and calculates the median vibration frequency of each target monitoring node according to the evaluation indicators. When the median vibration frequency shows a decreasing trend, it is determined that the operating state of the preset transformer is in an abnormal change state, and a fault warning is issued; otherwise, no determination is made.

[0071] A terminal, characterized in that it includes a processor, an input interface, an output interface and a memory, wherein the processor, input interface, output interface and memory are interconnected, wherein the memory is used to store a computer program, the computer program includes program instructions, and the processor is configured to call the program instructions to execute the method as described above.

[0072] A computer-readable storage medium is characterized in that the computer-readable storage medium stores a computer program, wherein the computer program includes program instructions, and when the program instructions are executed by a processor, the processor executes the method as described above.

[0073] In summary, the present invention utilizes the monitoring and analysis of voiceprint signals, combined with the technical means of multiple monitoring points, dynamic monitoring, and enhanced learning, to achieve early fault warning and accurate diagnosis of the transformer operating status, which is highly practical and innovative.

[0074] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A transformer core vibration monitoring and fault diagnosis method based on voiceprint features, characterized in that: The following steps are involved: Step S10: obtaining the voiceprint signal of the core vibration at each monitoring point in the preset transformer, collecting voiceprint features according to the voiceprint signals at different monitoring points, and analyzing the change rules of each voiceprint signal; Step S20: according to the change rule, the voiceprint signal change is distinguished into a slight change and a moderate change, and the operation state of the preset transformer in the case of the slight change is obtained; the operation state includes collecting the current parameters of the preset transformer; Step S30: According to the slight change given monitoring period, a critical threshold value of the change of the vibration frequency of the voiceprint signal is preset in different monitoring periods, and before the critical threshold value, the collected change of the vibration frequency of the voiceprint signal is divided into in, represents the vibration frequency of the nth voiceprint signal. In a certain monitoring period, if the vibration frequency of the voiceprint signal changes toward the critical threshold, it is determined that the change of the voiceprint signal evolves from a slight change to a moderate change, and a fault warning is issued for the preset transformer; Step S40: Based on the evolution law, analyzing the monitoring period in which the vibration frequency of the voiceprint signal with the most occurrences changes toward the critical threshold, and analyzing the change trend of the current parameter in the monitoring period; Step S50: presetting a safety threshold according to the change trend of the current parameter, and during the monitoring period, if the change of the current parameter does not exceed the safety threshold, then canceling the fault warning; otherwise, then not canceling it.

2. A transformer core vibration monitoring and fault diagnosis method based on voiceprint features as claimed in claim 1, characterized in that: In step S10, the monitoring points in the preset transformer include the surface of the core column, the inner wall of the transformer oil tank, the winding insulation pressure plate, the surface of the transformer oil tank, the insulation layer between the core and the winding, the bottom or top of the transformer oil tank and the core grounding lead.

3. The transformer core vibration monitoring and fault diagnosis method based on voiceprint features according to claim 1, characterized in that: In the step S30, according to the monitoring cycle given by the slight change, including monitoring with each natural day as a monitoring cycle, a set of initial vibration frequencies of the voiceprint signal is obtained in different monitoring cycles, and the number of the initial vibration frequencies is 3 to 5; the current parameters corresponding to each of the initial vibration frequencies are collected, and the current parameters are analyzed and compared according to the initial vibration frequencies obtained in different monitoring cycles, and the difference effect of different initial vibration frequencies on the current parameters is calculated, and the calculation is obtained according to the following formula: Among them, p j represents the pth initial vibration frequency obtained in the jth monitoring period; In the formula, k j represents the kth maximum current parameter collected in the jth monitoring cycle, x represents the average vibration frequency obtained from the initial vibration frequencies of the same group of voiceprint signals, h represents the difference in the initial vibration frequencies of the same group of voiceprint signals, and t represents the time required for adjacent differences to change; Within the time period, when the change of the current parameter is faster than the change of the adjacent difference, it is determined that the preset transformer is in an abnormal operating state and a fault warning is issued; otherwise, no determination is made.

4. A transformer core vibration monitoring and fault diagnosis method based on voiceprint features as claimed in claim 3, characterized in that: When it is determined that the preset transformer is in an abnormal operating state, a monitoring duration is given in the corresponding monitoring cycle, and the monitoring duration includes continuous monitoring for 10 to 15 minutes; 3 to 6 frequency values ​​with the highest vibration frequency are collected during the monitoring duration, and the frequency values ​​are marked as reference values; and the current parameter change is obtained based on the smallest frequency value, and when the current parameter does not increase with the increase of the frequency value, the fault warning is released; Otherwise, the fault warning is maintained.

5. A transformer core vibration monitoring and fault diagnosis method based on voiceprint features as claimed in claim 4, characterized in that: Collect the change trend of the average vibration frequency in different monitoring cycles of each of the monitoring points and generate a data set; obtain at least 3 monitoring points that change towards an increasing trend in the data set, and analyze the maximum current parameters of the corresponding monitoring points when the increasing trend changes, and preset a risk threshold according to the average vibration frequency corresponding to the maximum current parameter; when the change trend of the average vibration frequency of other monitoring points of the preset transformer in the future period is greater than that of the 3 monitoring points, it is determined that the preset transformer is in an abnormal operating state, and a fault warning is issued; Otherwise, no judgment is made.

6. A transformer core vibration monitoring and fault diagnosis method based on voiceprint features as claimed in claim 5, characterized in that: When it is not determined that the preset transformer is in an abnormal operating state, a preset acquisition time is set at the corresponding monitoring point, and the acquisition time includes one acquisition time per minute; within the acquisition time, the time required for each increase of 1 Hz frequency is obtained based on the average vibration frequency corresponding to the maximum current parameter. When in any of the monitoring cycles, if the time shows a shortening trend, it is determined that the operating state of the preset transformer is in an abnormal change state, and a fault warning is issued; Otherwise, no judgment is made.

7. The transformer core vibration monitoring and fault diagnosis method based on voiceprint features as claimed in claim 5, characterized in that: The Q-learning algorithm is used in the data set for reinforcement learning, wherein the reinforcement learning includes giving at least three target monitoring nodes in any monitoring cycle, wherein the three target monitoring nodes include a morning node, a noon node, and an evening node, and calculating the proportion of the vibration frequency at each target monitoring node to the vibration frequency of the corresponding monitoring cycle, which is calculated according to the following formula: Among them, w i represents the wth maximum vibration frequency obtained for the ith time in the monitoring period; Where y represents the maximum vibration frequency obtained at each target monitoring node, m represents the average vibration frequency obtained at each target monitoring node, (l i ) represents the lth vibration frequency difference collected at each target monitoring node for the i-th time; According to the preset proportion value of the vibration frequency of each target monitoring node in the corresponding monitoring period, when the vibration frequency of each target monitoring node is lower than the proportion value, it is determined that the operating state of the preset transformer is in an abnormal change state and a fault warning is issued; otherwise, no determination is made.

8. The transformer core vibration monitoring and fault diagnosis method based on voiceprint features as claimed in claim 7, characterized in that: Analyze the change law of the vibration frequency at each target monitoring node, divide each vibration frequency into a number of evaluation indicators, and calculate the median vibration frequency of each target monitoring node according to the evaluation indicators. When the median vibration frequency changes in a decreasing trend, it is determined that the operating state of the preset transformer is in an abnormal change state, and a fault warning is issued; Otherwise, no judgment is made.

9. A terminal, characterized in that: The method comprises a processor, an input interface, an output interface and a memory, wherein the processor, the input interface, the output interface and the memory are interconnected, wherein the memory is used to store a computer program, the computer program comprises program instructions, and the processor is configured to call the program instructions to execute the method according to any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, wherein the computer program includes program instructions. When the program instructions are executed by a processor, the processor is enabled to execute the method according to any one of claims 1 to 8.

Images